Polymer Melt Intercalation in Organically-Modified Layered Silicates

8000

Macromolecules 1997, 30, 8000-8009

Polymer Melt Intercalation in Organically-Modified Layered Silicates: Model Predictions and Experiment Richard A. Vaia*,† and Emmanuel P. Giannelis* Department of Materials Science and Engineering, Cornell University, Ithaca, New York 14853 Received March 5, 1996; Revised Manuscript Received September 16, 1997X

ABSTRACT: The effect of silicate functionalization, anneal temperature, polymer molecular weight, and constituent interactions on polymer melt intercalation of a variety of styrene-derivative polymers in alkylammonium-functionalized silicates is examined. Hybrid formation requires an optimal interlayer structure for the organically-modified layered silicate (OLS), with respect to the number per host area and size of the alkylammonium chains, as well as the presence of polar interactions between the OLS and polymer. From these observations and the qualitative predictions of the mean-field lattice-based model of polymer melt intercalation (preceding paper in this issue), general guidelines may be established for selecting potentially compatible polymer-OLS systems. The interlayer structure of the OLS should be optimized to maximize the configurational freedom of the functionalizing chains upon layer separation while maximizing potential interaction sites with the surface. The most successful polymers for intercalation exhibited polar character or contained Lewis-acid/base groups.

1. Introduction Polymer melt intercalation of mica-type layered silicates is a viable approach to synthesize a variety of polymer-silicate nanocomposites.1-6 These hybrids exhibit improved modulus,7-9 decreased thermal expansion coefficient,8 reduced gas permeability,8,10 increased solvent resistance,6 and enhanced ionic conductivity3,11 when compared to the pristine polymers. Furthermore, polymer-silicate nanocomposites, in addition to their potential applications, are unique model systems to study the structure and dynamics of polymers in confined environments.4,5,12,13 The nanoscale layered structure of the mica-type silicate leads to hybrids with nanoscale phase dimensions. Mica-type layered silicates, MTS’s, (alternatively referred to as 2:1 layered silicates) possess the same structural characteristics as the well-known minerals talc and mica.14-16 Their crystal structure consists of two-dimensional layers (thickness ) 0.95 nm) formed by fusing two silica tetrahedral sheets with an edgeshared octahedral sheet of either alumina or magnesia. Stacking of the layers leads to van der Waals gaps or galleries. The galleries (alternatively referred to as interlayers) are occupied by cations which balance the charge deficiency that is generated by isomorphous substitution within the layers (e.g. Al for Si or Mg for Al). In contrast to pristine mica-type silicates which contain alkali metal and alkali earth charge-balancing cations, organically-modified layered silicates (OLS) contain alkylammonium or alkylphosphonium cations.17,18 The presence of these organic-modifiers in the galleries renders the originally hydrophilic silicate surface organophilic. Depending on the functionality, packing density, and length of the organic modifiers, the OLS’s may be engineered to optimize their compatibility with a given polymer. Nanocomposite synthesis via polymer melt intercalation involves annealing, statically or under shear, a mixture of polymer and layered silicate above the softening point of the polymer. During the anneal, † Present address: Polymer Branch, Materials Directorate, Air Force Research Laboratory, Wright-Patterson AFB, OH 45433. X Abstract published in Advance ACS Abstracts, November 15, 1997.

S0024-9297(96)00348-8 CCC: $14.00

polymer chains diffuse from the bulk polymer melt into the interlayers or galleries between the silicate layers. Depending on the degree of penetration of the polymer into the MTS framework, hybrids are obtained with structures ranging from intercalated to exfoliated. Polymer penetration resulting in finite expansion of the silicate layers produces intercalated hybrids consisting of well-ordered multilayers with alternating polymer/ silicate layers and a repeat distance of a few nanometers. Extensive polymer penetration resulting in delamination of the silicate layers produces exfoliated hybrids consisting of individual nanometer-thick silicate layers suspended in a polymer matrix. In contrast, if the polymer and the silicate are immiscible, macro- instead of nanocomposites are formed. These macroscale hybrids consist of agglomerates of the layered silicate (sometimes as large as 1 mm in diameter) surrounded by polymer and resemble conventional filled rubbers and plastics. With organically-modified layered silicates, nanocomposites have been obtained from a large spectrum of polymers with varying degrees of polarity and chain rigidity, including polycarbonates, polymethacrylates, polysiloxanes, polyphosphazenes, and main-chain liquidcrystalline polymers.19 Whether an admixture of polymer and OLS produces an exfoliated or intercalated nanocomposite or a conventional macrocomposite depends critically upon the characteristics of the polymer and the OLS. These characteristics include the nature of the polymer as well as the type, packing density, and size of the organic modifiers on the silicate surface. Unfortunately, existing guidelines as to the optimum polymer-OLS combination have proven unsatisfactory and, at times, contradictory. Consequently, hybrid synthesis is currently a tedious trial-and-error process. In the preceding paper in this issue, we presented a mean-field, lattice-based description of polymer melt intercalation in organically-modified layered silicates.1 In general, an interplay of entropic and energetic factors determines the outcome of polymer melt intercalation. Free energy plots and their dependence on internal energy and entropy suggest three possible equilibrium statessimmiscible, intercalated, and exfoliatedsall of which have been experimentally observed. © 1997 American Chemical Society

Macromolecules, Vol. 30, No. 25, 1997

Polymer Melt Intercalation 8001

In the current paper, we report how various characteristics of the silicate and polymer, including silicate functionalization, anneal temperature, molecular weight of the polymer, and constituent interactions, affect hybrid formation. By comparing the experimental results with the theoretical predictions of the mean-field model, we then propose general guidelines governing hybrid formation.

The thermodynamic aspects of hybrid formation may be examined using the mean-field, lattice-based description of polymer melt intercalation as discussed previously.1 A great advantage of the model is the ability to analytically determine the effect on hybrid formation of various characteristics of the polymer and OLS. Briefly, the free energy change per interlayer volume, ∆fV, associated with polymer intercalation, may be expressed as

∆fV ) ∆eV - T∆sV

∆sV ) NAkB[∆sVchain + ∆sVpolymer]

{ ()

φ ˆ 1 π2 a1 φˆ 2 ln(c)(χS - χS0) ν2 ν1 6 h

abbreviation

M hW

PS30 PS90 PS400 PS3Br PVCH PVP

30 000 90 000 400 000 55 000 97 000 150 000

M h W/M h N Tg, °C 1.06 1.06 1.06 2.00 2.10

96 100 100 113 120 104

with γAB ) 2(γ+ describing the electron donor character, j j - 1/2 γj ) . Using the geometric combination rule, the total interfacial energy between species j and k is AB γjk ) γLW jk + γjk

(4)

LW AB where γjk and γjk are the polar and apolar components of the interactions, respectively, and

LW 2 - xγLW γLW jk ) (xγj k )

(5)

+ γAB jk ) 2(xγj - xγj - xγk )

(6)

These expressions provide a means to approximate the relative magnitude and sign of the pair/wise interaction parameters using available experimental values. Most importantly, they can be used to identify the range of interactions that is most favorable for hybrid formation.

2

)

2 1 2 ˆ 1φˆ 2 + ∆eV ) φ Q h0 sp,sa r2 ap

3. Experimental Methods

+

}]

a1 xm1 h

x3

(

poly(3-bromostyrene) poly(vinylcyclohexane) poly(2-vinylpyridine)

(1)

∆eV and ∆sV are the internal energy and entropy change per interlayer volume, respectively, and

[

polystyrene

γj ,

2. Thermodynamic Model

) NAkB

Table 1. Polymer Characteristics

u

(2)

(3)

∆sV is expressed as the sum of the entropy change associated with the organically-modified silicate, ∆schain , and the enV tropy change associated with the confinement of the polymer, . h0 and h are the initial and after polymer interca∆spolymer V lation gallery height, respectively. mi, vi, φˆ i, ri, and ai are the number of segments per chain, the molar volume per segment, the interlayer volume fraction, the radius of the interaction surface, and the segment length of the ith interlayer species. u is a dimensionless excluded volume parameter, Q is a constant near unity, and χS and χS0 are the fraction of interlayer volume near the surface at height h and h0, respectively, which influence the potential chain conformations. ap represents the pairwise interaction energy per area between the aliphatic chains and the polymer and sp,sa ) sp - sa, is the difference between the pairwise interaction energy per area between the aliphatic chain and the surface, sa, and that between the polymer and the surface, sp. A detailed discussion of the model including the different variables may be found in the preceding paper.1 Values for the geometrical parameters can be found in the general literature for a variety of polymers.20 In contrast, the as-defined values for the interaction parameters of an incompressible system are unavailable. However, since the model is mean-field, the microscopic pairwise interaction energies per area, jk may be expressed as an interfacial energy between species j and k, γjk. Then, ap ∼ γap and sp,sa ∼ ∆γs. The polar and apolar components of the constituent interactions may be separately examined using expressions for interfacial tension such as those developed by van Oss and co-workers.21-23 From their expressions,21-23 the surface energy, γj, is comprised of an apolar component, (γLW j , Lifschitz-van der Waals) and a polar/associative component (γAB j , Lewis acid/base). The Lifschitz-van der Waals component arises from dispersive and dipolar interactions, whereas the polar component arises from associative-type interactions. For the polar component, two parameters must be specified, one describing the electron acceptor character, γ+ j , and one

3.1. Materials and Synthesis. Organically-modified layered silicates were synthesized by a cation exchange reaction between the layered silicate host and excess alkylammonium cations, as outlined previously.18 Li+-fluorohectorite (F, exchange capacity ) 150 mequiv/100 g, Corning Inc.), Li+saponite (S, exchange capacity ) 100 mequiv/100 g, Southern Clay Products), and Na+-montmorillonite (M, exchange capacity ) 80 mequiv/100 g, Swy-1, University of Missouri Source Clay Repository) were used as received. Organic modifiers included dioctadecyldimethylammonium bromide (2C18, 2C18H37N+2CH3Br-, Kodak), octadecyltrimethylammonium bromide (Q18, C18H37N+(CH3)3Br-, Aldrich), and a series of primary alkylammonium chlorides (CnH2n+1NH3+Cl-, where n ) 6, 9-16, and 18). The primary alkylammonium cations were synthesized by acidifying a solution of the corresponding amine (Aldrich Chemical) with 1 M HCl. All OLS’s were dried in a vacuum oven at 100-130 °C and stored in a vacuum desiccator with P2O5 until use. The characteristics of the different polymers used in this study are summarized in Table 1. In general, hybrids were synthesized using the following procedure. Dry organosilicate (25 mg) and polymer powder (75 mg) were mechanically mixed and formed into a pellet using a hydraulic press at a pressure of 70 MPa. This insured the OLS was in the presence of excess polymer melt. Hybrid formation was accomplished by statically annealing the pellet in vacuum at temperatures greater than the glass transition temperature of the polymer and quenching in air to room temperature. Unless otherwise indicated, the samples were annealed to equilibrium, indicated by no further changes in the X-ray diffraction pattern. 3.2. Characterization. X-ray diffraction spectra were collected on a Scintag Inc. Θ-Θ diffractometer equipped with an intrinsic germanium detector system using either Cu KR or Cr KR radiation. After the raw diffraction data were corrected with the Lorenz-polarization powder factor and linear background subtraction, the location and width of the X-ray reflections were determined by profile fitting a Pearson VII function to the diffraction peak.24 All X-ray diffraction spectra were collected from samples that were quenched in air from the anneal temperature.

4. Results The outcome of an anneal was determined using X-ray diffraction measurements by monitoring the position,

8002 Vaia and Giannelis

Figure 1. Schematic depicting the expected X-ray diffraction patterns for various types of hybrid structures.

full-width-at-half-maximum (fwhm), and intensity of the (001) basal reflection of the OLS. Figure 1 summarizes the corresponding X-ray spectra for various types of nanocomposites. For immiscible polymer/OLS mixtures, the structure of the silicate is not affected by the anneal, and thus, the characteristics of the OLS basal reflection do not change. On the other hand, the finite layer expansion associated with intercalated structures results in a new basal reflection that corresponds to the larger gallery height of the intercalated hybrid. In contrast, the extensive layer separation associated with exfoliated structures disrupts the coherent layer stacking and results in a featureless diffraction pattern. The influence of polymer intercalation on the order of the OLS layers may be monitored by changes in the fwhm and intensity of the basal reflections. An increase in the degree of coherent layer stacking (i.e. a more ordered system) results in a relative decrease in the fwhm of the basal reflections upon hybrid formation. On the other hand, a decrease in the degree of coherent layer stacking (i.e. a more disordered system) results in peak broadening and intensity loss. 4.1. Packing Density, Ammonium Head Group, and Surface Charge. Table 2 summarizes the results for reactions carried out at 170 °C between PS30 and various octadecylammonium-modified OLS’s. The gallery height is calculated from the difference between the repeat distance, d001, and the thickness of a silicate layer, 0.95 nm.14-16 Annealing F18, FQ18, M2C18, and S2C18 with polystyrene yields intercalated hybrids. However, increasing the host area per octadecyl chain, as with M18 and S18, results in immiscible, unintercalated systems. Similarly, decreasing the host area per octadecyl chain, as with F2C18, also leads to an unintercalated system. At low interlayer packing densities of the organic modifier, the chains adopt a disordered monolayer arrangement. As the packing density increases, the chains adopt more extended conformations (and thus larger initial gallery heights), ultimately resulting in a solid-like paraffinic arrangement of the chains.17,18 Therefore, an intermediate range of host area per aliphatic chain and thus an intermediate range


of interlayer chain conformations are most favorable for polystyrene melt intercalation. The difference between primary and quaternary ammonium head groups does not appear to be a predominant factor for polystyrene intercalation. Both primary (F18) and quaternary (FQ18) octadecyl modifiers result in intercalated hybrids. However, the different type of ammonium head group does influence the net gallery height increase and the final gallery height of the intercalated hybridsthe net gallery height change for FQ18 is about 0.29 nm more than that for F18. The additional gallery height increase may reflect the larger size of the quaternary ammonium head group and the more delocalized cationic charge compared to that of the primary ammonium head group. In addition to the type of head group, the number of head groups per host area does not appear to be a determining factor. If the number of head groups was significant, the results in Table 2 should depend on the area per cation and not per chain. This is not the case. For example, M2C18 and M18 have the same number of head groups, but polystyrene intercalation occurs for M2C18 and not for M18. Furthermore, F18 has twice the number of ammonium groups as M2C18, but both OLS’s lead to intercalated hybrids. Finally, the results in Table 2 also indicate that the state of the polystyrene-OLS hybrid, whether intercalated or immiscible, is independent of the distribution of the anionic charge on the silicate surface. The negative charge of the silicate layers arises from isomorphous substitution in the crystal lattice. The location of the isomorphous substitution determines the distribution of excess negative charge on the layer surface and, therefore, the nature of guest interactions with the silicate surface.14-16 The increased localization of surface charge for tetrahedrally-substituted saponite does not appear to critically affect hybrid formation when compared to the more disperse surface charge of octahedrally-substituted montmorillonite. 4.2. Chain Length and Anneal Temperature. Figure 2 shows the final gallery height and full-widthat-half-maximum (fwhm) of a series of PS30-Fn samples annealed at 120, 140, and 160 °C; n is the number of carbon atoms of the aliphatic chain in the organosilicate. The gallery heights of pure Fn’s annealed at 160 °C are included for comparison. For n e 12, the gallery height of the unintercalated Fn is constant and the aliphatic chains are arranged in a pseudo-bilayer, with the highest segment density at n ) 12. For n > 12, the layer spacing increases and the chains adopt disordered pseudo-trilayer to liquid crystal-like arrangements.17,18 In general, the intercalation behavior of the Fn series is independent of anneal temperature between 120 and 160 °C but critically depends on the length of the chain, n. For pseudo-bilayer chain arrangements (n < 12), no polymer intercalation occurs. With the exception of a small increase in the fwhm of the peaks at 160 °C, the characteristics of the basal reflection, and thus the structure of the Fn’s, do not change. For n > 12, intercalated hybrids are formed, as evidenced by the gallery height increases. The full-width-at-half-maximum of the basal reflections of the intercalated structure is comparable to that of the original unintercalated silicate, indicating that polymer intercalation does not disrupt the stacking order of the silicate layers. For a full pseudo-bilayer (n ) 12) though, intercalated basal reflections are observed with intensities less than and breadths greater than those observed for the annealed



Table 2. Summary of Polystyrene (PS30) Melt Intercalation of Octadecylammonium-Modified OLS’s OLS

area/cation, nm2

area/chain,a nm2

initial gallery height,b nm

final gallery height,b nm

het change, nm

modelc

M18 S18 F18 FQ18 M2C18 S2C18 F2C18

0.72 0.58 0.39 0.39 0.72 0.58 0.39

0.72 0.58 0.39 0.39 0.36 0.29 0.20

0.75 0.83 1.33 1.57 1.43 1.50 2.85

0.75 0.83 2.16 2.69 2.25 2.35 2.85

0 0 0.83 1.12 0.82 0.85 0

intercalated intercalated intercalated intercalated immiscible

a Area per unit cell of aluminosilicate layer is 0.465 nm2.14-16 parameters in Table 4.

Figure 2. Final gallery height, h, and full-width-at-halfmaximum, fwhm, of a series of PS30-Fn samples annealed at 120 °C (g), 140 °C ()), and 160 °C (4) and pure Fn samples annealed at 160 °C (9).

Figure 3. X-ray diffraction patterns of a PS30-F12 sample before and after annealing at 160 °C in vacuum.

F12 host. Figure 3 compares a PS30-F12 mixture before annealing and after annealing at 160 °C. (The X-ray pattern of F12 for these conditions is independent of anneal time and temperature.) In contrast to the case for n > 12 and previous polystyrene-OLS hybrids, polymer intercalation of F12 disrupts the silicate layers, producing a more disordered intercalated hybrid. Increasing the anneal temperature to 180 °C alters the intercalation behavior of the Fn series for n e 12. Figure 4 summarizes the results for Fn and PS30-Fn samples annealed at 180 °C. Fn’s annealed in the absence of polymer at these conditions exhibit similar scattering behavior to that observed at lower anneal temperatures (Figure 2), indicating the Fn’s are thermally stable up to 180 °C. For PS30-Fn admixtures

b

Gallery height ) d001 - 0.95 nm. c Predicted hybrid states using

Figure 4. Final gallery height, h, and full-width-at-halfmaximum, fwhm, of a series of PS30-Fn samples annealed at 180 °C (O) and pure Fn samples annealed at 180 °C (b).

Figure 5. X-ray diffraction patterns of a PS30-F9 sample before and after annealing at 180 °C in vacuum.

with n > 12, intercalated hybrids form with gallery heights comparable to those formed at lower anneal temperatures. For all the bilayer chain arrangements (n e 12) however, the X-ray spectra contain broadened basal reflections corresponding to a disordered intercalated structure. For example, Figure 5 compares a PS30-F9 mixture before annealing and after annealing at 180 °C. The characteristics are similar to those displayed by PS30-F12 at the lower anneal temperatures. These results indicate that a bilayer chain arrangement and higher anneal temperatures favor the formation of hybrids in which the stacking of the silicate layers is disrupted. Because the samples are statically annealed, transport of the silicate layers over macro-


Macromolecules, Vol. 30, No. 25, 1997 Table 3. Summary of Melt Intercalation of M2C18 and F2C18 with Styrene-Derivative Polymers

Figure 6. X-ray diffraction patterns of PS30-F18, PS90F18, and PS400-F18 annealed 6 h in vacuum at 160 °C.

scopic distances to produce an exfoliated state would take a prohibitively long time. Therefore, this behavior may be indicative of a kinetically-limited exfoliated structure. 4.3. Polymer Molecular Weight. Figure 6 shows the X-ray diffraction patterns of PS30-F18, PS90-F18, and PS400-F18 after a 6 h anneal in vacuum at 160 °C. All three samples were prepared in an identical fashion using the same host particle size. The PS30F18 pattern contains diffraction peaks characteristic of only the intercalated hybrid. The PS90-F18 and PS400-F18 patterns though contain peaks characteristic of both intercalated (2θ ) 4.15 and 8.03°) and unintercalated OLS (2θ ) 2.82 and 5.66°). Fully intercalated hybrids for PS90 and PS400 are formed after annealing at 160 °C for 24 and greater than 48 h, respectively. The fwhm of the intercalated basal reflection is approximately the same for all molecular weights. Thus for statistically annealed PS samples, the final hybrid structure is independent of the molecular weight of the polymer. In this case, the molecular weight only appears to affect the kinetics of polymer intercalation. These observations agree with those of previous studies of the kinetics of polymer melt intercalation where the final hybrid structure was found to be independent of the polymer molecular weight as well as the particle size of constituents, the degree of powder mixing, and the processing (static or use of shear blending).4 However, the model1 indicates that there should be a slight molecular weight dependence for melt intercalation which arises from the entropy change of the polymer. Additional experimental work examining the intercalation behavior of a broader range of polymer molecular weights coupled with dynamic blending of constituents to enhance equilibrium mixing is required to further explore this issue. 4.4. Intermolecular Interactions. Table 3 summarizes the results of melt intercalation of a series of styrene-derivative polymers with M2C18 and F2C18. All the polymers are immiscible with F2C18. On the other hand, all polymers but poly(vinylcyclohexane), PVCH, form intercalated hybrids with M2C18, indicating that the structure of the pendant group on the polymer greatly affects hybrid formation.

a Intensity decreases observed associated with increased absorption of Br. b Peak broadening and intensity decrease of d001.

Figure 7. X-ray diffraction patterns of PS-M2C18, PVPM2C18, and PS3Br-M2C18 hybrids with the same volume percent M2C18 (20%).

Figure 7 compares the X-ray diffraction patterns of poly(vinylpyridine) (PVP), poly(3-bromostyrene) (PS3Br), and PS30 hybrids containing the same volume percent of M2C18. As discussed before, polystyrene produces a well-ordered intercalated hybrid. Bromination of the phenyl ring in the meta position (PS3Br) produces an intercalated hybrid similar to polystyrene but with a slightly larger gallery height. The intensity of the intercalated basal reflection, though, is reduced approximately 50% with respect to that of the unintercalated PS3Br-F18 mixture, but the fwhm of the intercalated reflection remains unchanged. The large decrease in the intensity of the basal reflection of PS3Br-M2C18 is partially due to the large absorption coefficient of the bromine atom that, when intercalated, decreases the overall scattering factor of the silicate/ polymer layer structure. Substituting the ortho carbon with nitrogen (PVP), produces a disordered intercalated structure that is similar to that observed for PS30-F12. The basal reflections of the PVP-M2C18 hybrid are substantially broader (0.019 rad) than those observed



may correspond to an intermediate intercalated structure. The intensity decrease of PS3Br-F12 is greater than that attributed to the increased absorption of the bromine atom seen for PS3Br-M2C18. Since the kinetics are slower for these polymers, the structures observed may be kinetically limited because of the static processing. 5. Discussion

Figure 8. X-ray diffraction patterns of PS3Br-F12, PVPF12, and pure F12.

for PS30-M2C18 (0.007 rad), indicating that the coherent length of the silicate layers in the PVP hybrid is less than that of those in the PS30 hybrid. Melt intercalation of M2C18 with these polymers emphasizes the importance of polar interactions during hybrid formation. Since the backbone of the styrenederivative polymers is comprised of a saturated hydrocarbon, polar interactions arise from the relative acid/ base character and degree of polarizability of the pendant ring. From small molecule analogs, the asymmetry of the electron density distribution of the pendant ring and thus the polarizability increases in the order PVCH 12 in Figures 2 and 4.31 When the gallery area available is greater than the lateral cross-sectional area of the chains, a pseudobilayer is formed. These structures are outside the boundaries of the mean-field model because their interlayer does not initially have ‘fully-occupied’ lattices.1,19 For the Fn series, interlayers not ‘fullyoccupied’ correspond to n < 12. Qualitatively though, as the area available per chain increases above the lateral cross-sectional area of the chain, the number of chain segments that gain conformational freedom as the layers separate decreases. Subsequently, the magnitude of ∆schain decreases. Additionally, the smaller V initial gallery height results in a larger penalty for polymer confinement. Hence, as with paraffin-type structures, hybrid formation should also be entropically less favorable for short chain lengths or interlayers with low chain-packing density. The intercalation behavior of PS-Fn’s with n < 12 as temperature increases is thus related to energetic changes in the system and not entropic changes. Combining these qualitative arguments with the results from the model indicates that, entropically, hybrid formation is most favorable for OLS’s with intermediate gallery heights and packing densities, in agreement with experimental results for the octadecylammonium-modified OLS’s in Table 2 and the Fn series in Figures 2 and 4. 5.2. Internal Energy. From the thermodynamic model, the energy change for hybrid formation (eq 3) depends on the sign and magnitude of the interaction parameters (jk) and on the initial interlayer structure of the OLS (Q, h0, and r2). In general, for small, favorable internal energy changes, intercalated states are expected. As the change in internal energy becomes more negative, the tendency to form exfoliated hybrids increases.1 We will first consider the effect of the interaction parameters on the internal energy change of the system. 5.2.1. Interaction Parameters. The interaction parameters, jk, express the relative energy change associated with the establishment of contacts between the polymer, aliphatic chains, and the interlayer surface. The interaction energy within the interlayer volume corresponds to the energy of polymer-aliphatic contacts, ap, whereas the interaction energy at the


interlayer surface corresponds to the replacement of aliphatic-surface contacts by polymer-surface contacts, sp,sa.32 Since the aliphatic chains are apolar (γ+ a ) γa ) 0), dispersive, van der Waals-type interactions will determine the sign and magnitude of ap. Equation 5 indicates that these interactions at best will be equal to zero for apolar polymers with surface energies similar to those of the aliphatic chains (γLW ∼ γLW a p ). If these dispersive interactions drove polymer intercalation, hybrid formation would be energetically most favorable for polyalkanes and saturated polyolefins. Experimentally though, these polymers lead to immiscible hybrids (PVCH (Table 3), polyethylene19) in static experiments.33 Thus, the favorable internal energy change driving polymer intercalation arises from interactions at the interlayer surface, sp,sa and ∆γs. If the polymer interacts more favorably with the surface than the aliphatic chain, sp,sa < 0. From eq 4 and 6, favorable interactions AB LW arise when γAB jk < 0 and |γjk | > γjk . Two conditions AB + + + exist for a negative γjk : γj > γk and γj < γk , or γj < + - 21-23 γk and γj > γk . Physically, these conditions imply that each interacting species possesses either the strongest electron donating or electron accepting moiety but not both. For the styrene-derivative polymers examined, these interactions most likely occur between the interlayer surface and the pendant groups of the polymers (Table 3). For polystyrene, the resonance structure of the phenyl ring exhibits a slight Lewis-base character.21 For PVP and PS3Br, the electronegative additions in the pendant ring redistribute the ring’s electron density. Possible Lewis-acid sites on the silicate include the cationic head group of the organic modifier or the Si4+ atoms in the tetrahedral sites of the silicate layer, while possible Lewis-base sites include the surface oxygens of the layer.14-16,35,36 Since to maximize compatibility monopolar (i.e. either Lewis acids or bases) components are required, as the basic character of the polymer increases, organosilicates with strong electron accepting and weak electron donating properties are needed. 5.2.2. Initial Interlayer Structure. In addition to the sign and magnitude of the interaction parameters, the internal energy change depends on the relative contribution of these interactions. The relative contribution is determined by the fraction of the various interaction sites within the gallery. This depends on the initial interlayer structure of the OLS. From the model, the relative number of the various interaction sites is expressed by the critical interlayer structure parameter, ξC, which is the ratio of the geometrical prefactors of the interaction parameters in eq 3.1 For large initial gallery heights (large ξC), the interactions within the interlayer between polymer and aliphatic segments, ap, contribute the most to the internal energy change. For OLS’s with smaller initial gallery heights (smaller ξC), the interactions at the interlayer surface, sp,sa, contribute increasingly more to the energy change. Recall that, for OLS’s with aliphatic chains, ap is greater than zero, and thus sp,sa must be less than zero for an energetic decrease upon hybrid formation. Note that the polar moieties of the polymer, which are necessary for the establishment of more favorable surface interactions, will be deleterious for interactions with the apolar aliphatic chains within the interlayer volume (eq 6). Thus, the optimal OLS interlayer structure will minimize the number of potential interac-


tions within the interlayer, while maximizing the number of potential replacement sites on the interlayer surface. A comment is warranted at this time concerning melt intercalation of layered silicates modified with organic chains possessing polar moieties such as phenyl, alcoholic, ester, etc. Proper choice of these moieties will result in more favorable interactions between the polymer and organic modifier, potentially leading to ap < 0 and very favorable conditions for melt intercalation. However, introduction of these moieties may also be deleterious. If these polar groups have very favorable interactions with the silicate surface, replacement of surface contacts will be difficult and the effective number of contacts between the polar groups and the polymer in the interlayer will decrease. Similarly, the entropic gain associated with the increase in conformational freedom of the organic chains will decrease. These effects will impede hybrid formation. A careful weighing of all these considerations is necessary to select the optimal interlayer functionalization. Qualifying the influence of the initial interlayer structure and the various interlayer interactions on hybrid formation may lead to a criterion to select potentially successful polymer-OLS systems. One potential approach to determining the equilibrium state for a polymer/OLS system is to calculate the free energy change per area using eq 1-6. Using values in Table 4, we determined the equilibrium states for various FnPS systems (12 e n e 18) from the minimum of the free energy change per area as a function of gallery height.1 In all cases intercalated hybrids are predicted in agreement with the experiment. The final gallery heights predicted from the model for different chain lengths also exhibit the same trend (decreasing intercalated gallery height with decreasing n) found experimentally (Figures 2 and 4). Within the framework of the model, the effect of varying chain length parallels that for varying the area per chain, since both contribute linearly to h0 (eq 7). Therefore, assuming the interaction parameters are the same for the different OLS’s, the hybrid state predicted for OLS’s with various areas per chain (Table 2) also agrees with the equilibrium states observed. However, the magnitude of the net gallery height increase calculated from the model differs from experimental values by up to 35%. This is not surprising though given the simplicity of the theoretical formulation and the uncertainty of the values of the parameters used. 5.3. Product Maps. An alternative approach to using the free energy curves to determine the favored equilibrium state is to derive a ‘product map’ (similar to a phase diagram) for a given set of parameters. The product maps display the dependence of the sign of the internal energy change on two system parameters. As outlined in the previous paper,1 the total entropy change, to a first approximation, is near zero for the initial stages of polymer intercalation. For these cases, the sign of the internal energy change will determine the sign of the free energy change of the system and, thus, the outcome of intercalation. Within the bounds of the model, the total energy is near zero when |sp,sa/ ap| ) ξC.1 Determining ξC from the initial interlayer structure and using eqs 4-6 to express sp,sa and ap, the resulting equality gives the dependence of the point of zero energy change on the various apolar and polar interaction parameters of the constituents. From the perspective of establishing a methodology for selecting


Figure 10. Product map showing the dependence of the internal energy change per interlayer volume on the Lewis LW ) acid, γ+ p , and Lewis base, γp , parameters of a polymer (γp 42 mJ m-2) with an ‘octadecyl’-modified silicate. Curves are determined using the surface tension parameters in Table 4 and ξC ) 5.0.

successful polymer/OLS combinations, the product maps are more instructive than the free energy curves because the effects of various parameters on the equilibrium state are more readily discernible. For example, using parameters for the OLS outlined in Table 4, Figure 10 shows the relationship between the Lewis-acid, γ+ p , and Lewis-base, γp , components of the polymer interfacial energy for intercalation of a -2 polymer (γLW p ) 42 mJ m ) into an ‘octadecyl’-modified silicate. Similar maps may be constructed displaying the relationships between other interaction parameters, such as intercalation of a given polymer into modified silicates with different γ+ s and γs . In the central region of Figure 10, the internal energy change, and thus the free energy change of the system, is less than zero, and hybrid formation is favorable. For polymers with parameters outside this region, the net energy change is greater than zero and hybrid formation is unfavorable. In principle, the potential product of polymer melt processing may be determined a priori from values of the surface interaction energies. Because of the limited availability of measured values of LW γ+ j , γj , and γj , a comprehensive comparison with experiment is difficult. However, for the values available, the product maps agree well with the experiments.19 Currently, we are exploring ways to better quantify the various interaction parameters. 6. Conclusions Experimental results indicate that the outcome of polymer intercalation depends critically on silicate functionalization and constituent interactions. We observed that (a) an optimal interlayer structure of the OLS, with respect to the number per area and size of the alkylammonium chains, is most favorable for hybrid formation and (b) polymer intercalation depends on the existence of polar interactions between the OLS and the polymer. In general, the conclusions of the mean-field model described previously1 agree with the experimental results. The entropy loss associated with confinement of a polymer melt is not prohibitive to hybrid formation


because an entropy gain associated with layer separation balances the entropy loss of polymer intercalation, resulting in a net entropy change near zero. Thus, from the theoretical model, the outcome of hybrid formation via polymer melt intercalation depends on energetic factors which may be determined from the surface energies of the polymer and OLS. From these observations and the construction of product maps, general guidelines may be established for selecting potentially compatible polymer-OLS systems. Initially, the interlayer structure of the OLS should be optimized to maximize the configurational freedom of the functionalizing chains upon layer separation and to maximize potential interaction sites at the interlayer surface. For these systems, the optimal structure appears to exhibit a chain arrangement slightly greater than a pseudo-bilayer. Polar polymers containing groups capable of associative-type interactions, such as Lewis-acid/base interactions or hydrogenbonding, lead to intercalation. The greater the polarizability or hydrophilicity of the polymer, the shorter the functionalizing groups in the OLS should be to minimize unfavorable interactions between the aliphatic chains and the polymer. Acknowledgment. This work was supported by NSF (DMR-9424446) and by generous gifts from DuPont, Exxon, Hercules, Monsanto, Nanocor, Southern Clay Products, and Xerox. R.A.V. gratefully acknowledges the partial support of a NDSEG Fellowship. We would like to thank L. Vega and R. Krishnamoorti for fruitful discussions. This study benefited from the use of Material Science Central Facilities at Cornell. References and Notes (1) Vaia, R. A.; Giannelis, E. P. Macromolecules 1997, 30, 7990. (2) Vaia, R. A.; Ishii, H.; Giannelis, E. G. Chem. Mater. 1993, 5, 1694. (3) Vaia, R. A.; Vasudevan, S.; Krawiec, W.; Scanlon, L. G.; Giannelis, E. P. Adv. Mater. 1995, 7, 154. (4) Vaia, R. A.; Jandt, K. D.; Kramer, E. J.; Giannelis, E. P. Macromolecules 1995, 28, 8080. (5) Vaia, R. A.; Jandt, K. D.; Kramer, E. J.; Giannelis, E. P. Chem. Mater., in press. (6) Burnside, S. D.; Giannelis, E. P. Chem. Mater. 1995, 7, 1597. (7) Usuki, A.; et al. J. Mater. Res. 1993, 8, 1179. (8) Yano, K.; Usuki, A.; Kurauchi, T.; Kamigaito, O. J. Polym. Sci., Part A: Polym. Chem. 1993, 31, 2493. (9) Messersmith, P. B.; Giannelis, E. P. Chem. Mater. 1994, 6, 1719. (10) Messersmith, P. B.; Giannelis, E. P. J. Polym. Sci., Part A: Polym. Chem. 1995, 33, 1047. (11) Aranda, P.; Ruiz-Hitzky, E. Chem. Mater. 1992, 4, 1395. (12) Vaia, R. A.; Sauer, B.; Giannelis, E. P. Submitted to J. Polym. Sci., Part B: Polym. Phys. 1997, 35, 59. (13) Wong, S.; Vasudevan, S.; Vaia, R. A.; Giannelis, E. P.; Zax, D. J. Am. Chem. Soc. 1995, 117, 7568. (14) Pinnavaia, T. J. Science 1983, 220, 365. (15) Brindley, S. W., Brown, G., Eds. Crystal Structure of Clay Minerals and their X-ray Diffraction; Mineralogical Society: London, 1980. (16) Gu¨ven, N. In Hydrous Phyllosilicates; Bailey, S. W., Ed.; Reviews in Mineralogy Vol. 19; Mineralogical Society of America: Washington, DC, 1988; pp 497-560. (17) For a detailed discussion of the interlayer structure of alkylmodified OLS’s, see ref 18 and references contained therein. (18) Vaia, R. A.; Teukolsky, R. K.; Giannelis, E. P. Chem. Mater. 1994, 6, 1017. (19) Vaia, R. A. Doctoral Thesis, Cornell University, 1995. (20) Brandrup, J., Immergut, E. H., Eds. Polymer Handbook; John Wiley and Sons: New York, 1989. (21) van Oss, C. J. Interfacial Forces in Aqueous Media; Marcel Dekker: New York, 1994. (22) Norris, J.; Giese, R. F.; van Oss, C. J.; Costanzo, D. M. Clays Clay Miner. 1992, 40, 327.

Macromolecules, Vol. 30, No. 25, 1997 (23) van Oss, C. J.; Chaudhury, M. K.; Good, R. J. Sep. Sci. Technol. 1989, 24, 15. (24) Brown, A.; Linde, S. Adv. X-ray Anal. 1986, 30, 343. (25) The asymmetric electron density of the ring moieties was approximated from 1H NMR spectra of cyclohexane, toluene, pyridine, and 3-bromotoluene by comparing the relative shielding and deshielding of the ring protons. (26) In contrast to the planar conformation of the rings with resonance structures, cyclohexane posses a boat conformation, and thus it might interact differently with the OLS. For a first approximation, we ignore these geometric effects. (27) Experimental results reported by van Opstal et al.28 for mixtures of polystyrene and short n-alkanes indicate that the upper critical miscibility temperature (USCT), and therefore the Θ condition, of these mixtures are near temperatures used for melt intercalation of polystyrene. This agrees with ap ∼ 0 determined from the surface tensions, lending validity to our methodology of calculating values for the interaction parameters. (28) van Opstal, L.; Konigsveld, R.; Kleintjens, L. A. Macromolecules 1991, 24, 161. (29) Barton, A. F. M. Handbook of Solubility Parameters and Other Cohesion Parameters; CRC Press: Boca Raton, FL, 1983. (30) Israelachvili, J. Intermolecular and Surface Forces; Academic Press: New York, 1992. (31) The length of a fully-extended alkylammonium chain may be approximated as (n′ - 1)0.127 + A + B nm where n′ is the number of carbon atoms and A and B are the size of an ammonium and a methyl group.14-16

Polymer Melt Intercalation 8009 (32) It may be argued that the polymer-OLS interlayer interactions are associated with dipole-induced interactions resulting from electrostatic fields arising from the anionically-charged silicate surface and the cationic organic modifiers. Assuming the cation head groups reside at the silicate surface,19 the interlayer E B field may be shown to be approximately zero using Gausses law. This is a reasonable assumption, since the cationic head group would prefer to reside close to the anionically-charged surface. (33) Values for γLW are not available for some of the polymers p used in the study. However, the dispersive interactions between the polymer and the aliphatic chains may be estimated by examining the difference in the dispersive cohesion energy densities, δd, of polymers and alkanes. From 29 and δalkane ) 16.6) 15,34 δPS the literature, δPVCH d d ) 19.7, d 16.8 MPa1/2.29 The most favorable apolar interaction occurs when the difference in dd’s is zero. If dispersive interactions within the interlayer drove polymer intercalation, PVCH would be more favorable than PS. Experimentally, this is not the case (Table 3). Furthermore, polyethylene (δPE d ) 17.6 MPa1/2 29) would be even more favorable than both PS and PVCH. However, melt intercalation attempts between linear polyethylene and OLS’s have been unsuccessful.19 (34) Gehlsen, M. D.; Bates, F. S. Macromolecules 1994, 27, 3611. (35) Bleam, W. F.; Hoffman, R. Inorg. Chem. 1988, 27, 3180. (36) Bleam, W. F.; Hoffman, R. Phys. Chem. Miner. 1988, 15, 398.

MA9603488

Polymer Melt Intercalation in Organically-Modified Layered Silicates

Recommend Documents